<p>This study aims to develop an efficient English pronunciation assessment system based on interactive security design and deep learning technology, in order to improve learners’ pronunciation ability while ensuring user data security. This system adopts deep learning speech recognition algorithms for speech signal preprocessing to improve data quality; designed a deep neural network model structure and optimized recognition accuracy through reasonable parameter selection. The overall framework of the system is designed around user interaction experience, implementing statistical parameter speech synthesis methods and pronunciation optimization algorithms to achieve effective evaluation of user pronunciation. To ensure the interactive security design of the system, an identity authentication module was researched and designed, which adopts a multiple authentication mechanism to ensure the security of user information. The application of network security situational awareness algorithms enables the system to monitor potential security threats in real-time. By optimizing the secure storage speed of files, the secure access of user data is ensured, further improving the system’s interactive performance. Therefore, the English pronunciation evaluation system based on deep learning not only achieves efficient pronunciation evaluation function, but also ensures the security of user data through interactive security design, providing safe and convenient support for future English learning.</p>